two & a half gamers session #23 - How to do a proper game performance audit? UA, Admon & GD

[Music] hello hello hello this is no gaming podcast to half gamer session number 23 we're discussing the latest interesting news and fun knowledge let's not forget this is still a 4:00 a.m. conference discussion Vibe so let's not take it too seriously today we have Felix covid catcher braber in the house thir times the charm yeah yeah exactly Yak microphone hugger REM and myself mate FAL boy L mate like I'm thinking about starting a business where basically I sneeze in envelopes of people who want to have covid cuz I always have Co like this is like and speaking of 400 a.m. discussions you were awake at 4:00 A.M today weren't you no today but a day

before so yeah we had very interesting conversation at the festival obviously I met a lot of um former colleagues actually it's kind of nice it's kind of nice how are you besides Co obviously I feel like I feel like as well Remo come on cheer us up how do you feel yeah pretty good like I would say you know not not drinking not doing festivals everything's great this week not uh like heat waves s uh declined so that's good like nice 23 25 degrees I'm so worried about your immune system man because like it never gets exposed it never gets punched in the gut like mine does you know like it's not hardened you know I have all the vaccines man me too I got CO as well even you know even after three shots of uh what it was astroica no F yeah kind fiser now

I have what they call like hybrid immunity but I guess Felix you you don't have that anymore or maybe or maybe you have now even better immunity for at least at least like one month I don't know man I've so many goddamn times right now so like you know I don't even think I have immunity or I'm actually think my body is incapable of having immunity for this this is like third or fourth time it's so many times I've actually lost count like it's yeah what can you do I mean I'm expecting Co I have Co like three times until the end of the year most probably you never know you never know yeah okay so um today the skip news but we will definitely have fun and drop some knowledge so last session with our super important friend Eric sparked one uh one topic that was in my head and it's always in my head when it comes to starting to work with the company so how

to do a proper game performance audit and since we have uh ad monetization game design in UA in uh in here let's talk about that so I think can we actually start with you Remo because you always speak and it's Never Ending Story So I think uh if we want to keep keep Pace let's start with you and then uh we can mix it up a little bit okay sure we can do that um yeah I've done this a few times already when we were all uh working on the other side of the barricade pretty much so this is just like a short version of my usual checklist that I would do uh but yeah let me start that I would cut these into three categories that is the team the production and the game itself which I think all deserve like a special kind of attention themselves so let's start with the first one which is the most important I think is the team so usually when you go and

look into like a new game or you know you're doing a partnership or maybe you're doing an acquisition or something kind of a similar or just helping someone uh first thing would be definitely to check the team and their past experience uh we talk about this many times here but genre expertise is really invaluable like if you know you have people that have been making I don't know match three games for 10 years and now they're making RPG I think something doesn't click there but again on the other hand if you have let's say a team that's been I don't know switched from Peak into dream games and now he's making pretty much a similar game you know that they know what they're doing so looking for genre expertise in the past experience is really important and pretty much taking is some kind of a risk factor because what I do and like how this kind of a list that I do is pretty much you're trying to check out the red flags and know that like yeah look here and look under this mattress and look under this you know cabinet and take out pretty

much all the skeletons so you know what you are kind of looking at uh the other important thing when looking at the team is to get a clear view of who is actually designing and leading the game because this is kind of a hazy or let's say not really clear thing in our industry sometimes games are driven by just product managers sometimes designers are somewhere there at the you know edge of the production doing their designs and then just kind of giving them to programmers so it's really important to get some kind of a strong feel of who is actually leading the project who is actually deciding like this feature will go out this feature won't go out this feature is I don't know too big for us this feature will cut down stuff like that and then when you when you actually pinpoint the person then comes the other other part of the team discussion and you are trying to pretty much discover how up to dat are again with the genre Innovation so you talk with them about their you know their vision for the game the current Trends especially

competitors that's really important like you need to find out if they actually know their competitors and I know it's not like I mean know doesn't mean that they know you know the names of their competitors but actually how their competitors work how they work means how do they make money which features are the critical ones that they use in their like feature portfolio if they are you know innovating on those if they're just like cloning something and pretty much getting that kind of overall feel of like how how the team is designing the game and what their USP that's also one of the kind of tricky points here when lots of times I would know that the team is just kind of cloning something and maybe doing some kind of really really small Innovation to it but the big question is do they really know what like if this will work if they have a clear plan that like yeah we're going to capture this audience and this audience going to be you know taken by us because we have this other feature which going to Delight the users because we're better than this one and or we're innovating on the previous game stuff like that so I guess T would be the team

and the last one that a little bit skips already to the production category is that awareness of the team regarding their feature road map how long some features take to to be actually created and implemented because sometimes and I know this for many times I've seen this in the industry that people who have these kind of really really ambitious road maps and like yeah we're going to do matchmaking and PVP and clans feature and social feature and they jam it into like two months and then you look at it and like no way way no way everything everything looks good on paper yeah paper can take it that's the problem so so this is again really really critical point to kind of find out if people know how much time it will take them to do some of those features and if the team like did them before or have any experience with them they allocated enough capacity because usually what happens that road maps as they go further and further starts to crumble and like yeah we cannot do this we cannot do that we late two months we got this one cut that one and so on so

for so it's really important to see if they're in syn with all of this so if you see this in the in the road map like what what do you do what's the what's the plan afterwards like what's the next step for the the team to depending depending on what's the result of this but let's say if if it's too ambitious you try to cut it down to something that testable because the most important thing is to get some kind of actionable kpis as soon as possible so you know you're not building something that won't work so and you don't need to wait like six months until you get everything in place and and you know the important thing is that have stuff that's kind of actionable or seen by the user at least until day three if you are building you know that really really far ahead Clan feature or whatever meta Game feature that will be I don't know from Level 58 that the team is already working on it it won't you know make the cut like if if and every game fails if they cannot achieve day one retention in the end so you try to make it great granular enough

so they can pretty much go and take one Milestone after the other that that's that's really important and yeah that goes into the we're switching again into the second category which is the production and yeah that's the first question is the road num realistic uh do you see the capacity of the team being able to actually Cod up with it uh road map is there enough programmers are there enough backend people how's the you know even the state of the code I I know like I cannot really you know discern this or like see it out like how it is how good it is but it's really important to know that the code is okay because I've seen teams fail because their code based went so bad that just updating it took them so much time that they pretty much craw to Halt how do you check this it's quite subjective right not not the code part but like yeah you need for this but U sometimes you can if you talk to people symptoms kind of arise that that will tell you again these red flags that maybe something is not really okay which means like yeah we're updating the build once per month

really like for like something that's in production that's like too slow and like why are you updating so you know so infre infrequently or some kind of other kind of a red flags that arise if you talk about this and from that point on you can determine that there are some problems so Yi mentioned this on the on the interview we we did together in the in the past like he is also like evaluating the companies based on like discussions and questions so he's asking a lot of questions and then trying to check these red flags in between the lines and like in the in the answers and then yeah like it's pretty much it's pretty much the same like you know like a doctor's appointment like you tell him his symptoms or you try to pretty much ask enough questions so you get the symptoms and from the symptoms you okay is this disease is this other disease or like there's you know nothing there's a problem uh but you know sometimes the doctor actually do the checkup and then try to ex I mean like what's what's wrong not only ask questions yeah yeah exactly sometimes you need to get your hands dirty and like you know go into

documents read the documents read the Excel sheets read the dashboards or stuff like that if there are even any and do it yourself because this whole process is just to kind of cut down all the red flags that you don't know about so you're trying to pretty much make sure that okay you look into this and this and this and there's no problem there uh last thing regarding production is of course soft launch and the road map for soft lunch this is same category as the normal road map whereas the team will tell you like okay we're going to do this and that and then soft launch for three months and then Global launch like really month no no no like better is like okay so we have soft we will have soft Lun one month because everything's ready like oh who wait wa wait a noks or anything and you're like man like you really want to test this before you do Global launch mate sure don't worry you know we were in the pre-production for like six months and like everything is there so we have the game ready it's going to fly like yeah of course okay when someone tells everything is there there usually

there's something missing that's a red flag that's usually red FL instead of them asking like is everything there or like them saying like yeah it's all there then you're like yeah something's missing I guarantee it yeah like like let's take some industrial examples for instance I don't know brawl star like brawl star was in soft launch for one and a half years and they changed the meta game they they repaired now it's making millions of of dollars and supercell they can afford being in soft lunch one but imagine how how much supercell has and even though they has it it took them one and a half years to actually execute on all those things so if you know if if like oh we have this problem with monetization oh we're going to be figure out in soft lunch in three months like yeah sure yeah definitely so again it's about the dialogue with the team and their kind expectations it's about management of expectations and you try to you know rule them into something realistic and if uh if you are repairing these already you try to allocate time again into some

kind of actionable steps make the wh flunch granular in a way like okay we don't have any data yet let's just focus on the day one retention so we know the core gameplay works because day one retention is usually this the you know D derivative of your core gameplay and then goes the other day three day seven day 30 and then your monetization and so on so forth those all these other Milestones but if you focus on all of these at once you know capable of everything master of nothing so I I wouldn't really recommend do do that uh last thing is the game itself and this goes even before like talking all these other questions pretty much do your homework before how long is the game in production who started the game and if he's still there we didn't team because this is a really really like frequent occasion whereas you come you come into the team you ask them how's the game going how he been going and then suddenly you learn out that the person who actually was the creative brains of the game left out already like a year ago or half a year ago or three months go so this is really important because

then it means that somebody else is either doing a pivot or they are following into steps of someone there's not there to kind of help them direct uh next thing would be again expectations regarding kpis as the team itself if they have any kind of a expectations from their soft lunch kpis if they're Aiming High or aiming low or if they're aiming even anywhere like worse worst answer you can get like how do you think about your monetizations we want to make so much money so that's really bad but actually people knowing that I don't know we're in this genre and this genre is really really LTV hungry because the CPI is really high so we need to know and like adjust the features that based on that stuff like I don't know ads we know this genre will be 7030 so 70% IPS 30% ads we want to implement this and that and that for that we have this system in the game economy that's allocated into the budget of the at revenue and the IP revenue and stuff like that and again just by talking about these things you can clearly see where the team is heading

what's their current status and what's their plan and their execution uh again this just goes into more and more details such as like knowing the competitors kpis at least you know like everybody knows that I don't know merge games currently has like over 20% or 15% day 30 retention because they're so addictive and like if you're aiming for I don't know 5% day 30 or currently have 5% day 30 like you know like you won't make it stuff like that uh last thing is again is there any plan behind the economy and monetization uh is there a person that's kind of managing the sinks and sources if it's balanced at least on paper because even if you balance it on paper you still need to check it with dashboards and analytics if it's actually you know doing what you want it to do but in the end there's a clear plan which means somebody is purposely creating the economy to create you know deficiencies so at some point you will hit the wall and need to pay or just you know grind out or kind of a plan

somewhere ads yeah watch a couple ads also this is really also important because usually if there are economies run on ads which means that I don't idle economies or these other like w reward in the end beautiful economies run on ads yeah these need to be built from the startup from the startup you need to know like okay people will be watching ads and doubling the rewards meaning that the double rewards are not bonus they are the default so if you're not doubling them you're crawling and very important point you made like I can't stress how important that is like that's such a good point why you just you you can just duct tape the ads on top of everything yeah we have an economy and let let's just double the income that will be great of course yeah inflation is only in real life not in game economies yeah yeah come on yeah crypto people would have their input here definitely game games are not games are not real come on of course infation doesn't exist there yeah and the last thing which is I guess the most advanced question from the list would be liveops

uh because this is again really really common problem where the whole team is doing this kind of Heroic Production heroic soft lunch initiative they finally make it they finally you know Global launch the game spend tons of UA budget and then suddenly there's no content after three months and people are leaving like crazy because they haven't thought about liveops and additional content threadmills that you know these games are gam as Services those are not like finish and move to next one so there needs to be some plan with some kind of a Content threadmill or whatever kind of a Content that you're creating that your team is capable of delivering after global lunch keeping the players in and of course your golden cohort like how the golden cohort is going to eat through your content because those are payers they will eat through the very very fast and you need to have some kind of backups in order to keep them there because once they eat through all the content they leave if there's nothing else to eat for them of course they will do it super fast so if you are expecting to have a content for three months it's gonna be

maybe one and a half month maybe one month for them yeah like having having some kind of let's say new event one month or even during that first month of global launch is completely okay and like you're not over stretching it you're actually keeping people in so definitely have that some some kind of a I would always advise have this kind of recurring event template to have at least that that you can recycle at least when the team kind of hops on that content threadmill so you have at least this kind of a recurring content that you maybe you know change skins change visual but at least there's something new being updated into the game all the time because if there's like a three months without an update people are leaving like and mass so that there's a big problem yeah I think that's it uh regarding the really short version uh what's your team production only took 18 minutes well done man yeah well done this is the short version man this short version usually takes like two days [Music] two days of you talking of course well

that's uh yeah that's how we know you it's like uh yeah okay but I can continue with the UA and and creatives basically um I usually look at different things and just to give you an idea of what I'm looking at it's uh it's Channel performance then creative uh performance in history and then I'm trying to to look at Daily spend versus daily revenue what kind of percentage here the game has from Paid UA and Organics a percentage of installs of course and doing this data health checks because well you can run a lot of UA activities and um from time to time it's very important to check the data because nine out of 10 companies that I work with uh and starting to work with they don't have the data in order and and from time to time you see very big discrepancies between different uh UA sources or well not the UA sources but

data sources if you compare MMP um Facebook dashboard or um Google Play console or or apple console you see three different numbers and that's very that's that's very bad what's the cause usually like you know it's it can be so there can be like very uh wrong implementation of the SDK you can look at at the different attribution windows which is uh not a bad thing but you need to have that in mind but usually um an issue with an implementation of sdks you know like from time to time even when when you do these like new updates so sometimes programmers is like oh well this there's this SDK nobody uses it and uh or he or she didn't ask anyone like hey is this actually something that you use and you manager would SC like yeah of course we need this but then sudden sometimes can this can happen SDK

disappears and then you can't uh track purchases or something like that happens and it's I mean it's quite common yeah it's very common very common and even within vicinity to 20% yeah just just the myth itself yeah but um when I'm looking at these discrepancies um if I'm seeing 5 to 10% maybe 15% differences that's fine I mean 20 is like okay that's a a little bit of red flag but then like everything um above 20% it's just yeah well stop stop UA don't waste money or well at least make this a priority because we need to check that yeah getting back to the to the actual UA and the channel performance so checking all the all the channels that are live that's that's the first thing to check and then everything that was um tested and worked before and trying to find like what worked what

didn't and why and I'm usually looking at the campaign structure and the campaign type so let's take for example Facebook so um are is's take Google man stupid Facebook Facebook it's not good anymore let's do a good one well I I will yeah don't worry I will I will speak about I speak about Google as well uh you know based on the campaign type is the is the company running the Tria campaigns just automated at ad app ads or is it a regular type of campaign the old one where you can actually Target multiple things like interests or look likes uh are they running the the campaign budget optimization are they running purchase campaigns or value optimized campaigns also like this is also tied to what we discussed um a couple of episodes before is important to check like what is the monetization strategy of the game if like you said

there's 70% inaps versus 30% ads so is it uh are they running only um in a purchase campaigns or add campaigns as well that's important if it's if it's 50/50 or even 7030 can can work you need to mix and match the the multiple optimization structures and then I'm usually looking at uh like what kind of GEOS uh the game is running which is also important because I like to to actually Target the whole world but you need to think about like different Geo um buckets usually there's like us stand low sometimes I put us in the T1 bucket because uh it tends to work well when uh you have multiple DRS around the US and then most companies are actually afraid of running tier two or tier three countries not true why but if you mix it up efficiently you can you can see pretty good results from tier two tier

two countries and tier three countries because the CPI obviously the cpis are not that high and the LTV is not that high as well but if you still can can have a a healthy margin why would you why wouldn't you running right so usually using these like tier tier um country buckets or uh Regional targeting which is um well available everywhere even on Google unity and uh Tik Tok and Facebook and then uh it's also important to check what kind of age segments um you are targeting sometimes people just do like um everybody and uh then if you Dr drill down the Ed segments you will see like the best results for example for 25 to 45 and like everything else is just not efficient spent so you can cut that and immediately see some kind of bump in terms of the numbers and then since you know uh Android devices are kind of well

there's a lot of them right so you can yeah you can you can have really high-end devices and also very shitty devices and um and what I usually do uh when I see the Android uh is not performing that well then I'm trying to check uh what kind of devices uh have what kind of performance and do some kind of exclusion list so back in days uh there was a study I think not V did it or someone else um they tried to check the Performance Based on the ram instead of um the actual operational system because you can have uh the latest operational system and then the the device is pretty shitty anyway because you bought it for 30 $30 so so usually do these like exclusion list based on RAM and everything that don't have like at least like two or three gigabytes of RAM I just try to exclude it because that's

like that's a sign of a highend device at least for me I mean you can you can do even four gab of ram these days and get pretty high devices what are the you know most usual red flags in the channel performance typ well uh for example only one type of uh optimization so let's say they're running only value campaigns or only purchase campaigns or they're running only certain amount of GEOS so us and tier one or only English English speaking countries like why so small small diversification yeah yeah yeah yeah uh even like within the one channel or okay they're running only one or two channels not not uh not others and when I ask like why like how well it didn't perform like when did you try like hey two years ago oh yeah okay all right that's uh you know that's a different industry in our case yeah yeah yeah two years ago I mean well a

lot of changes happened so like this type I also I also try to ask questions because usually when I do these like audits I don't speak to the team before because I don't want to hear their um opinions and like and they and don't don't want to ask questions because I just want to see the data and then I ask why why you do why you did this why you did that and then the answer because I just don't want to have biased opinion before doing that audit it's kind of nice because then I have all the questions it's good which is good and um I mean that's like also within the Google there is uh you can use different um types of creatives different again segments um Geo segments and also different optimization methods so uh that's based on like Channel performance but then a lot of uh

I put a lot of emphasis on the creatives and um I usually try to look at like what type of oft creatives are running at the moment if it's static images videos playables videos and playable and cards and and different uh different things I usually try to keep like all the all the types of creatives um in the mix even on Facebook running static images plus videos well playable is not that much because on Facebook it doesn't make any sense to to run run playable unfortunately I know Facebook pushed it super hard but then man the performance is terrible and um in terms of the that networks I definitely try to push for playables because there's uh where where it makes sense and uh and also on Google static Imes plus videos I usually static images um and different I I take a step back the different types of creatives on Google you will get different types of

inventory so if you if you go with um images you can get pretty shitty Google Display Network performance and a lot of scam and then if you if you use um a lot of videos different formats and different lengths of videos you can get a pretty high end um inventory from YouTube that's also yeah sorry one question regarding fakeness how do you approach that one within the creatives if they're using it or not or like should they be using it or do you propose that like they should be using it well it depends on the game and uh and performance of the creative they have and what worked and what didn't work uh before it's not necessary always to push for for fake fake ads um in also like I tried to see if they used it before and how did they look like because you can have fake which is super fake and uh it won't work because well it doesn't have anything think in common with the game play MH but you can have

fake uh that is actually kind of connected to the game play so like dream games fake like dream games fake yeah exactly so I don't always push for for fakes but you know sometimes uh you need to be brave enough and crazy enough to to try different uh different Creative Concepts then I Tred to look at like top five and the bottom five creatives uh you know because viously it's important to see what worked but it's also really important to see what didn't and then trying to understand like why it could not perform or what why it could it could perform and also talking to the creative team and uh and trying to understand like what's their process for production for Creative production because it's also very important sometimes you have like this super long processes in big corporations and sometimes you have these L processes even within the small teams like and I try to ask like why why

the would you do these like long processes super easy just brainstorm look at the competition brainstorm again then produce test evaluate and then the whole and repeat yeah exactly that's it so uh so that's but it's also like again a dialogue with the team so looking at the data and then talking to the team to understand what are they doing how do they approach the UA how they approach the the channel testing and also creative testing as well and then as I said like I'm trying to understand like what's the daily spent what's the daily revenue what kind of goals uh the UA team have because UA team has because there are these like different team different Ras goals on the day seven obviously I don't think it makes a lot of sense but okay whatever I still think like you need to have a benchmarks on day one day seven day 30 and some times even on a day 14 or even day three it's

like if you have a goal on like day seven then you don't care on what happens after day seven and that's that's what I've seen before for um within some UA teams because you know like okay we we hit the day seven goal but you didn't you know like the the LTV didn't grow after day seven I don't care I at the goal so just set set me a different goal oh okay well that's it's a interesting approach so uh so yeah I think uh that's uh that's briefly as well on the UA side cool thanks M I guess it's over to me and Co Casa uh CC yeah Casa so over to admon so mine I think is like the most nonsubjective thing we've talked about cuz it's all in the numbers like everything without is just beautiful it's right there so what

happens first is like um first what you do is uh you I used to open up the mediation dashboard and what I'd like to say the first thing I do is a sanity check so I look at first uh the historical Revenue curves of the game going back about two and two to three months you're looking for like any Peaks or troughs that can't be explained by any holidays or any like other events if you see any like Peaks or troughs that don't really make sense you should really look into why that was then you should break down the same time period take the revenue break it down by Network and look here again for any Peaks or like troughs and usually like you can see if there's a big drop you can even break it by placement and you can try to localize which placement is broken or which one is seeing like demand shifting also kind of the sanity check part you should look at is you look at you break also The Impressions over Impressions per Dau and add viewer rates and basically get a sense if your like

user base behaviors are shifting if so like you should find out why the golden metrics you're kind of looking for in terms of Impressions per da for rewarded anything above five is considered very high uh then also kind of you can shift into looking at the ad viewer rate so the percentage of people who watch ads in your game so for rewarded anything below 45% I would say is quite bad and basically that is an indicator that your placements are not good enough or that your rewards are not sufficient enough for people to watch it anything above 60 for rewarded is considered very very good so yeah uh you can even like on iron Source mediation you can break it down further and you can look at the completion rate that is to say what is the portion of users that finished a rewarded ad that started so if it's lower than around 92% that's an indication that the rewards that you're

giving out for this placement are not high enough to Warrant users to complete watching the ad so that is an indication so if you have high like ad view rate but low completion rates that means that your rewards are not high enough and that you should increase the rewards the next thing you should look at is are the sdks up to date so SDK is like their pain in the ass to update I know development teams like hate it we yeah but like overall performance always increases when you like improve the SDK and download a newer like version so there is a trade-off here right so usually what I say to development teams is when an SDK comes out like don't jump head first into it like on the first day like Yay you don't want to do that so I'd say the ideal time to install a new SDK is one to two weeks after it being live because sometimes when these SDK go live happens maybe like three three four times a year if you're working with a couple of OD networks they will crash

your game and screw everything up so what I always say is let other people test the SDK stability for you and then after a week if there's no complaints or emails coming out from the ad Network that's when you install it and let other people kind of be the guinea pigs you shouldn't be the guinea pig um so moving on to Phill so that's the the next thing after doing all the sanity checks that I look at hang on I'm going to a little Co in my throat um so moving on to Phil so I can't stress this enough like Phil today like there is no excuse for a game to have below 977 to 98% fill on whatever like ad unit they're using for pretty much any goo like even if you have you know a couple million impressions a day in India you will get that filled at One Price probably with aobi because they're

really strong in India like you can localize this and if I see that there's any like any fill that's currently below 97% that's just unacceptable like uh if you have that like that means that you need to add more ad networks and basically maybe move the placements down but there is like yeah no excuse even putting in a back fill placement that fills at any like cost that's what you should be doing in Iron source it's really easy to check uh like what the fill rate is in the apps but in app loving it's a bit harder to check overall fill I've been told that you should basically look at the app loving fill as a proxy for overall fill so whatever the app loving bidder is filling at that's a good proxy what overall fill is in the app um yeah say what you want with that so then the next thing I look at is ecpm so here's a bit harder to say what is good and what is bad but rule of thumb here is that an average good which

is like right on the money like that's like right in the middle of yeah this like split is that worldwide ecpm on rewarded video should be around 15 and USA it should be around 30 so anything that's above or below those two numbers that's either bad or good so for high organic games this is always lower because organic traffic tends to be basically not priced as high as in organic traffic and the reason why is because when you do paid UA uh essentially networks have like tagged that user so they're aware of what that user is worth in their own network so let's say if Facebook like brings like a user to your game and if they notice that like they're willing to bid more on that user because they know that that user has behaved in a certain way before so that's kind of the rule of thumb for

ecpm um is there is there sorry is there any way to like kind of increase it if I'm that organically heavily dependent game like I don't know Subway Surfers or something like that yeah so usually ad networks favor like uh what can you say locations or games that have a lot like a lot of Impressions so if you have you can increase the Impressions per da or overall impressions in the game that generally will increase your ecpm you also but if you think about it ecpm is just a function of like how the you your users performed with that particular Network in the last seven days like CTR install rate uh like all those things so there's nothing really you can do to improve ecpm and maybe you shouldn't either because like you got to remember like you get paid more if more users leave your game right yeah like a good click through ra anything between like se7 to 1.2% that's like decent click-through rate but like yeah nothing I can say off the bat okay the next

thing here moving on to checking is revenue so Revenue you got to remember is Phill and ecpm put together and that's what you should always be optimizing for so here uh what you I always look for is the classic admon mistake that you always make in your career at least once and that is to see if the like mediation um like Revenue matches up with the dashboard Revenue so this is really quick to check in app Ling you can just do a network comparison that like checks it really like checks the Delta between the input like amount and like the actual amount you're getting paid out in Iron Source I believe you have to check it like in each individual like ad placement so what I'm looking for here is if I have a placement in unity that's $5 and in unity it's put as 15 then basically there's a disc connect and like I've had this a couple times where

like one time with a client I was just logged in and they were making like eight grand a day and everyone's super happy like how can we optimize this and then I look in and like the placements are wrong and they're really making a grand and a half and I'm just like hey guys hate to be the bear of bad news but you're not making eight you're making 1.5 like sorry and then you know there's the meslo hierarchy of needs like conversation that comes after that that begins with anger which is horrible to deal with yeah uh the next thing I look at which is yeah kind of the second last thing I even look at is the share of Revenue or share of Voice or wallet or however you say it uh so what I do is in the last 30 days I break down the ad like the ad Revenue by networks in the last 30 days you take the revenue by Network and you divided by the overall Network to like overall Revenue to look at the percentages what you're looking for here here is in a perfect world you want an even split among as many

networks as possible so if a network has more than 50% share of Revenue that's a yellow flag not a red flag but a yellow flag essentially that means that that network has price setting power in your app over your inventory because basically they're taking up majority of the fill so if you have that situation where maybe like a nap loving or iron Source or like yeah admob is taking up like more than half or almost half you need to try to get that down and when you try to get that down you will improve competition in your app and you will improve overall Revenue like it's a deadly trick so usually the best way to bring that down is like yeah be more aggressive with optimizing placements adding new networks setting price floors doing that kind of things also one thing that people forget you can do here is like if you're already working with all the networks you think you can work with speak to the network AMS that you have and say like hey can we open up some new

demand sources so basically look at like the last 30 days of which like inventory has been pre predominantly displayed in your app and be like hey these genres are working really well on like the ad Network that's 50% can you force fill a couple of these campaigns just to see how it performs I think it'll perform well and usually like those things work so you got to remember like these smaller networks like kind of yeah the ones like the vles are like kind of the ones that are a bit smaller usually in shares of Revenue you can usually like ask them to for like yeah help you out and like with new demand sources and that's a good way of getting like their share Revenue up okay so you want to prevent the Monopoly pretty much exactly exactly and then uh the last thing I look at if I'm doing a health check or like kind of one of these checks is then take a quick look at the waterfall like does every placement have a different network between them so the issue is let's say like if you have Unity at yeah

like 50 60 70 80 90 with no network like between them or no network between them filling even what will happen is after a few weeks the unity bidder will figure out that like I don't need to pay 90 I don't need to pay 80 I don't even need to pay 70 nor 60 I can pay at 50 because there's nothing in between here so they look at the ad requests or the bidder looks automatically at the ad requests if it's roughly the same at the same amount like subtracted by The Fill then basically they're going to pay you less because there's no competition in between you broke the sandwich rule exactly like you need a sandwich yeah you need a sandwich uh yeah and then basically the last thing you look at the top placements like how much are they filling so you're looking for a high number here that's another yellow flag so like if you our top ad placement is like let's say fiber that's filling at $90 at % then you're like oh I need to add more like higher placements because you're willing to pay a lot more and that's basically what you look

for sounds pretty comprehensive it's very comprehensive and since I've been bedridden for this week and this is probably the most fun I've had in four days nice well nice any any last comments tips or or tricks that that we can share with with the guys and the audience yeah I would say regarding mine it's like mine is I guess the most stable one because I guess your guys is much more depending on the context and seasonality and like where are we within markets mine is more kind of prone to design Trends let's say or business model changes but they don't change that fast but yeah I would say regarding those two they need to be put into the context as much they said like yeah we tried two two years ago yeah like situation is completely different now so take it with the grain of salt if you're listening today or in one year ago two years yeah yeah yeah I have one question for you Ma

just off the cuff right now like in order from 1 to five what are the strongest UA like networks or SL sources you can use right now that what you're seeing just like if R rank them like one is Google two is Facebook like up to five well one would be Google then Tik Tok Unity uh I would say Facebook and maybe up I mean from the best to well not the worst but least efficient I don't see but like no one else I've ever heard in a podcast is that clear like no one says that like there something that yeah there something works at least for me but again uh this is something that works for me I would think it would work for you as well but you never know that's why I always uh try to recommend like a something like testing uh process or testing methodology get like 10 to 15% out of

your budget on a monthly basis and use that for testing you never know like uh what's changed within the platform if we had this like podcast recording like last year I wouldn't defin I wouldn't definitely say like Tik Tok is on second place I would say Facebook Google unity and that's it but Tik Tok definitely improved quite a lot in the in the last months Unity as well it was really good performing uh Channel last year there also changes um on like on the platforms them so definitely different the UA mix on Google than it's on iOS at the moment so yeah think about that cool nice sweet yeah I think there there you have it guys uh well don't forget don't forget to hit the Subscribe button because that's super important also for us but also for you because then you will get the fresh recording as soon as it's

uploaded also we're super Vain and we need those clicks yeah exactly all right perfect thank you very much guys see you see you next time [Music]